Forecasting realized range volatility: a regime-switching approach

Abstract

In this study, we introduce the regime changes to the heterogeneous autoregressive realized range-based volatility model (HAR-RRV). The different regimes are specified depending on the expectations of volatility of RRV volatility. Our out-of-sample findings indicate that considering regime switching can significantly improve the forecasting accuracy of volatility models.

Keywords:

• realized range-based volatility
• Markov regime switching
• forecast

JEL Classification:

• C22
• C52
• C53

Additional information

Funding

This work was supported by the National Natural Science Foundation of China [grant numbers 71372109, 71371157 and 71401077]; the doctoral program of higher education fund special research project [grant number 20120184110020]; the young scholar fund of science & technology department of Sichuan province [grant number 2015JQO010]; and Southwest Jiaotong University graduate student innovation experiment practice project [YC201405118] is gratefully acknowledged.